2.0 Analysis 2.1 Reconstruction of Events Leading to the Fire The reconstruction of events leading to the fire was based on the examination and testing of material evidence as recovered from the fire site, on the evidence as presented by the tunnelmen, in conjunction with on-site observations by personnel after the fire, and on the examination of photographic evidence taken shortly after the fire. 2.2 Reason for the No. 4 Gate Being Non-operational At the time of the occurrence, the hydraulic system was being tested for the second time. The vibrators were found to be functioning during the first test; this would indicate that the hydraulic system was under pressure, thus eliminating the possibility of air lock in the main hydraulic line. Further, as the TSB Engineering Branch tests conducted on the No. 4 gate valve (in use at the time of the fire) showed the valve to be functioning, the reason for the non- actuation of the No. 4 gate cylinders for a period of some two to three minutes before the fire cannot be explained. 2.3 Hydraulic Hose Disconnection and Initiation of Fire The fact that the hose was found disconnected after the fire and that the coupling threads were not damaged despite the high pressure within the main hydraulic line suggests that the hose may have been backed off manually and was barely held in position by the threads either with or without possible manual support from the head tunnelman (see Appendix C, Figure 2). Further, with the system pressurized, oil should have sprayed from the coupling the moment the coupling had been loosened. Although the hydraulic system was under pressure, there was a time lag of about two minutes before the pressurized oil was seen to come out from the disconnected hose. Thus, the hydraulic pressure had not reached the No. 4 valve until this stage. Tunnelman No. 1 reported hearing a weak noise that sounded similar to a vibrator, but different, a couple of seconds before he observed a large orange fire-ball. This is consistent with the hydraulic hose striking against metallic surfaces while spraying fluid in all directions because the hose which was under pressure was disconnected. The instantaneous hydraulic mist so generated then ignited when it came in contact with the halogen lamp which culminated in a substantial fire-ball. 2.4 Significance of the Flame Observed by Tunnelman No. 2 The flame sighted by tunnelman No. 2 and described as having been associated with burning would suggest that the flame was elongated in shape and similar to that produced by oxyacetylene equipment. However, this flame could not have been from the torch which was used in the repairs as the torch was in the starboard tunnel and out of his line of vision. Hence, any acetylene-fed flame had to have been from a minor leak in some other part of the equipment. In the absence of corroborating evidence to substantiate an acetylene leak in the tunnel prior to the onset of the fire and/or a source of ignition, the appearance of the flame described as having been associated with oxyacetylene equipment cannot be explained. Given the rapidity with which the fire progressed and in the absence of evidence to the contrary, it would appear that the flame sighted by tunnelman No. 2, thought to be in the vicinity of the cargo hold, was most likely the initiation of the hydraulic oil fire. 2.5 Victim's Position and Burn Pattern The head tunnelman's last communication with tunnelman No. 1 questioned whether the gate cylinder had moved yet which would suggest that he was preoccupied with the testing of the No. 4 gate. From that position, when the fire erupted, the halogen lamp would have been on his right side and could account for the heavier burns he sustained on that side, and the heavier burns to his chest would suggest that he was facing the fire-ball. As neither of the two surviving tunnelmen had observed the head tunnelman's movements immediately before and following the onset of the major fire/fire-ball, the precise reason for the position of the victim's body, some distance from the work site, cannot be ascertained. However, given the circumstances of the case and in the absence of evidence to the contrary, one possible explanation could be that as the head tunnelman was in the vicinity of the fire-ball, its effect would have thrust him in a forward direction and could account for the dislodgement of the boots and the head trauma sustained. The eruption of the fire would have ignited the victim's coveralls drenched in hydraulic oil, accounting for the burn pattern found on his body. 2.6 Progress of Fire As indicated by the tests, hydraulic oil poured on the surface of the halogen lamp could not have ignited, but the mist could. Further, the absence of the glass cover further facilitated the ignition of fire. By the time the hydraulic system was shut off, some 400 to 450 L of hydraulic oil had been dispersed into the tunnel space providing a large quantity of flammable oil. As the fire progressed, it was further fuelled by the oxygen and acetylene gases from the gas-cutting equipment, setting the conveyor belt on fire. 2.7 Approach to Fight the Fire One of the techniques used to fight fire is to starve the fire of oxygen. This is achieved by restricting and/or sealing off ventilation. In this instance, the ventilation to the tunnel was not in use. The presence of propellants in the tunnel space and the intense heat generated by the fire necessitated the opening of two cargo hatches to facilitate cooling the saddleback and flooding of the tunnel. This proved effective in preventing the fire from spreading. Further, the heavy smoke generated by the burning hydraulic oil and conveyor belt would have created a smothering effect on the fire. Thus, the cumulative effect of the cooling of adjacent metal surfaces and the smothering effect of the smoke generated resulted in effectively extinguishing the fire within a short time frame. The smothering effect from the smoke was diminished, to a certain extent, by the release of oxygen from the oxygen cylinder once the hose was burnt out. 2.8 Oxyacetylene Cylinder Because the torch was being used in the starboard tunnel while the cylinders were positioned in the centre tunnel and the closest connection between the two tunnels was at the entrance to the forward exit, the crew, of necessity, was required to walk a considerable distance to shut off the oxyacetylene valves. As repairs are physically demanding, time-consuming and sometimes frustrating, the crew set practices to facilitate operations. As the air flow in the tunnel was restricted (mechanical ventilation to compartment not used), any gas/vapour introduced in the tunnel would accumulate creating a potential fire hazard. There is a potential for generating hydraulic oil mist during the hydraulic system test. The oxyacetylene cylinder valves should be closed prior to and during such tests. In this instance, while the closing of the oxyacetylene cylinder valves prior to the hydraulic system test may not have prevented the onset of fire, the narrow flammable range of the hydraulic oil would probably have made the fire difficult to sustain for any significant period of time. Additionally, the fire would have been of shorter duration and of reduced intensity than in this occurrence. In view of the hazard associated with the use of oxyacetylene equipment in spaces other than those exposed to open atmosphere, the safe working practices call for compartments where oxyacetylene equipment is used to be well ventilated. In this instance, although the tunnel space was provided with forced air ventilation, it had not been utilized when oxyacetylene equipment was in use. Owing to the limitations of the fire-fighting equipment (including the absence of fire hydrants and fire-hoses) in the tunnel space, additional safety precautions would have been in order. As the nearest fire hydrant was in another compartment some distance away, a total of four hoses had to be connected to reach the fire site. Further, with the tunnel space limited and confined, handling fire-hoses under pressure was difficult and cumbersome with increased possibility for personnel injury. Hence, as a precautionary measure, a fire-hose with a combination nozzle should have been rigged, kept ready for use with water in the line, and a hydrant valve left open on deck. 3.0 Conclusions 3.1 Findings The head tunnelman was in the vicinity of the No. 4 gate valve at the onset of fire. The hydraulic hoses and fittings used were suitable for use, and there was no evidence to indicate that the hydraulic hose had burst. The hydraulic hose to the No. 4 gate valve was found disconnected after the fire. The cause for the disconnection is unknown. The hose which was under pressure became disconnected, allowing hydraulic oil to be sprayed forming a flammable mist in the tunnel. The hydraulic oil mist was ignited by the bare halogen lamp. A flame observed in the vicinity of the work area by the tunnelman was the initiation of the hydraulic oil fire. The halogen lamp fixture, which was intended for outdoor use only, was being used in the tunnel. The protective lens cover of the halogen lamp fixture was not in place and the illuminated lamp surface temperature was in excess of 600C. The lens temperature of the type of halogen lamp in use was in excess of the temperature permissible for handling lighting fixtures. The acetylene cylinder in use had replaced an empty cylinder earlier on the day of the occurrence. The acetylene cylinders, being part of the ship's stores, were not required to meet Canadian standards and, consequently, were not required to be fitted with fusible plugs, and they were not so fitted. There is no requirement for the oxyacetylene cylinders to be fitted with flashback arresters or for the hoses to have non-return valves, and they were not so fitted. The valves to the oxyacetylene cylinders were not shut prior to testing of the hydraulic system. The polyester-blend coveralls worn by the victim were less fire-resistant than coveralls made from pure cotton. The hazards associated with the flammability of hydraulic oil mist and the presence of a bare high-intensity halogen lamp were not fully appreciated. The victim lost consciousness due to head trauma and lost his life due to asphyxiation and extensive burns. Difficulties were experienced in handling long lengths of fire-hoses which had to be led from the laundry compartment. 3.2 Causes The fire broke out in the conveyor belt tunnel aboard the HALIFAX when the hydraulic oil mist from the disconnected hose was ignited by a halogen lamp that was missing a protective lens cover. 4.0 Safety Action 4.1 Action Taken 4.1.1 High-intensity Lamps Subsequent to the occurrence, the TSB apprised the Canadian Coast Guard (CCG) of the results of its laboratory tests and of the potential fire hazard associated with the installation and use of high-intensity lamps in an indoor environment aboard vessels. The CCG issued Ship Safety Bulletin (SSB) No. 1/95, in January 1995, advising all vessel operators and maintenance personnel of the following: the need for compliance with Ship Safety Electrical Standards (TP 127, Section 18(4)) concerning the use of portable equipment; the need for maintenance of all portable equipment in a safe working condition; use of portable shipboard equipment in a safe environment aboard; and the need for enclosure of fixtures in areas where there is a risk of spray or readily combustible material. 4.1.2 Application and Enforcement of Safety Regulations After reviewing comments from persons with a direct interest in the findings, the TSB apprised the CCG of discrepancies in the application and enforcement of safety regulations by regional offices. 4.1.3 Hot Work Procedures As a result of this occurrence, the owners have reportedly introduced a new procedure for conducting hot work aboard vessels. The hot work permit system is intended to ensure that all safety precautions are taken prior to the commencement of burning, welding, and other hot work using oxyacetylene gas.